Nikon D300 - weird metering behaviour

I have noticed the a strange behaviour in the D300 exposure metering and thought of picking the collective wisdom of the forum to see if there's simple solution.

As some of you know, I have a serious perversion towards shooting wreckscapes in the murky Northern Baltic waters. In our conditions (and especially when using strong filters) exposures tend to run in the tens of seconds. Historically, I have used manual mode for my shots but last night I felt lazy and selected aperture priority in stead.

Now, the correct exposure should have been in the 25-30 second range @ f11 (aperture selected to keep the exposure under 30 sec) and ISO 800. But - and this is an important word so I repeat it: But - the camera for some reason elected for 15-20 sec, resulting in grossly underexposed pictures.

I tested this further at home by putting a lens cap on and shooting in aperture priority and program mode and guess what, the metering was all over the place. Basically, when in aperture priority, the meter shows f11 and LO but the camera exposes for anything between 3 sec and 15 sec with no apparent change in anything. And with two different lenses too.

So I called my father who has a D200 and he tested the same and got more or less the same results. So it's a feature. And I do not like it.

I seem to remember my old D70 in similar conditions always exposed for the max 30 sec and then gave up. Which was infinitely better than an erratic behaviour. So my question is, is this metering behaviour for example configurable somewhere? The manual was not very helpful in this respect.

In the end, I have no problem shooting manual and will soon end up rigging a remote release to go Bulb for the deeper stuff anyway, but somehow I feel that paying good money for all this fancy software shoud give me something in return.

Puzzled,

timo

To continue,

It has been suggested that I'm getting some light onto the meter through the viewfinder. In film (TTL) cameras this would be the problem (and Nikon even supplies the little plastic cap that you lose within weeks of purchasing the camera), but my understanding was that in (most) digital SLRs, there's no metering going on when the shutter is open. Is this correct also for the D300?

Of course, any light leaking in through the viewfinder is bad and I suppose I need to rig a plug on the housing viewfinder. If the housing-camera joint was lightproof (which I doubt it is) this would stop any light leaking in. Oh, well, at least there will be less of it....

timo

Timo,

My guess is that because the light is so low and the resolution of the meter is such that it can't get an accurate reading. What I mean is that the difference in the amount of light at ISO 800 and f/11 between 15 and 30 seconds is such that the camera is having problems determining the difference. i.e. you are at the limit of the sensor.

Think of it like estimating the size of something. If something is around 1mm in length, you're not going to be very accurate in your estimate and you could be out by up to 100%. But if you see something that is 1 metre long, you'll be able to estimate it within 10%.

I just did an experiment with my D300 with slightly more light. I mounted it on a tripod in a dimly lit room. Camera was on ISO 200 and aperture priority. The first test was with the lens at f/14. The camera said I needed 20s, but would jump to 25s every now and then. I took 4 shots, 3 were at 20s and one at 25s. I then opened up the lens to f/10 and the camera said 10s was required. I took 3 shots and all were at 10s.

I'll try some more tests when I get a chance at higher ISO and under darker conditions, but I'm pretty sure that the darker it gets, the less accurate (and consistent) the metering will be.

Timo - it's the viewfinder option mate.

I have been playing on land with IR and the D300 - exactly the same behaviour ( IR is also crap on the D300 BTW .

You can try it out on land also - look at the exposure meter on the top of the camera with a stopped down daylight shot and cover the rear viewfinder - meter is all over the place. Seems they are calibrated with the expectation of your eye blocking the back.

The other issue is you can get some weird flare like blobs appearing on the shots from light entry into the rear viewfinder. I have seen that with quite strong sunlight at my back ( not a problem at depth I will wager ).

Paul C

Timo - it's the viewfinder option mate.

Thanks Paul,

That would explain it, especially when I just realised that although I do frame the image with my eye in the finder (live view, unfortunately, is pretty much useless in the dark), the metering actually takes place with the finder exposed. Now I just need to rig a cap to the finder. Fortunately the Hugy housing has a nice round protrusion where the finder exits the housing so this should be reasonably straightforward to remedy. At least to the point where it should not be a significant problem any more.

Overall, though, I have to say that the D300 performs admirably in these conditions. Compared to the D70(s) the high ISO noise is very low indeed, although starts to be really visible at ISO800 and above.

The image below was taken in bright sunny conditions in about 20 meters of water and 4-5 meters of viz using the Nikon 10.5mm. Exposure was 30 sec@f11, ISO400 and using in-camera NR. Filtration is Kodak Wratten Magenta 70cc (stacked 50+20) to get the contrast way up in the green water. The anchor shank is actually longer than the naked eye could see in the visibility of the day. I have not done any significant editing to this one, just set exposure and WB and tweak the contrast a bit.

So I think after 18 months of trial and - mostly - error, I'm starting to get somewhere.

timo

Interesting Stuff Timo,

Can you email me the RAW ?

I promise to sell it on if I can as my own work - oops - no, but I would be interested especially as the IR stuff has taught me a few tricks that I would like to try on UW images.

4m Vis ! Oh joy we have yet to get more than about 1m here locally. The frogs have been dredging some estuary and dumping the result at sea or due to a nasty weather front that went through in March (depends on who you believe).

The D300 is pretty good ( yea I have one also ). Will be trying it 'proper' next week as am off on a US shark thing with Alex.

Take care,

Paul C

Well, there you go. I wouldn't have thought that much gets in through the viewfinder, but I did a test (same as before) but this time shone a torch in through the viewfinder. The metered exposure time went from 13s to less than 1s.

I have an eyecup on my D300 and I leave it on when the camera goes into the housing (Ikelite). That greatly reduces (if not eliminates) any stray light getting in between the viewfinder port and the camera's viewfinder.

Can you email me the RAW ?

Will do. It lives in my home 'puter (homputer?), but I'll try to remember to send it when I get there tonight.

timo

Well, there you go. I wouldn't have thought that much gets in through the viewfinder, but I did a test (same as before) but this time shone a torch in through the viewfinder. The metered exposure time went from 13s to less than 1s.

Particularly important if you have a great thick filter on the lens!

Alex

Well, there you go. I wouldn't have thought that much gets in through the viewfinder, but I did a test (same as before) but this time shone a torch in through the viewfinder. The metered exposure time went from 13s to less than 1s.

 

I have an eyecup on my D300 and I leave it on when the camera goes into the housing (Ikelite). That greatly reduces (if not eliminates) any stray light getting in between the viewfinder port and the camera's viewfinder.

Yup - it surprised me as well - while playing with IR and exposures in the 30 sec region.

Borrowed a mates D70 in the end exposure went from 30 sec to 1 sec for the same F stop - man those Nikon guys and girls have tightened the response of the filter on the front of the CCD !

As for the stray light in Timo's case it may not help much as he is using a Tripod and I guess a timer start exposure. Push the button and two seconds later the shot kicks off. Meanwhile he will be moving back and stray light is coming in.

Pretty sure the exposure time is calculated just before the shot is taken - don't think that it is dynamic and calculated 'on the fly' during exposure.

The internal flare that you can generate is secondary but annoying if you have a bright ambient and a long exposure - killed a few of my IR shots

<edit> Hey Timo - if you gaffer tape over the viewfinder on the D300 BEFORE you put it in the housing she would be fixed a treat! - just a 'helpful' thought.....

Paul C

As for the stray light in Timo's case it may not help much as he is using a Tripod and I guess a timer start exposure. Push the button and two seconds later the shot kicks off. Meanwhile he will be moving back and stray light is coming in.

 

Pretty sure the exposure time is calculated just before the shot is taken - don't think that it is dynamic and calculated 'on the fly' during exposure.

Yep, I'm using a 10 sec timer. So the process is: setup tripod, aim, squeeze the nipple, move back and relax. I generally do not keep my eye on the finder while triggering in order to avoid shaking the tripod so the finder is exposed to ambient during metering regardless of when that happens during the process.

Going forward I will set up a wired remote (I have a third bulkhead already fitted) using an on/off switch to allow me to do Bulb exposures. I'm using fixed focus (as the AF would be useless anyway) so a simple flip switch does the job just fine.

<edit> Hey Timo - if you gaffer tape over the viewfinder on the D300 BEFORE you put it in the housing she would be fixed a treat! - just a 'helpful' thought.....

 

Paul C

Good idea

But joking aside, I had naively hoped that live view would have helped me in this as aiming through a 2+ stop filter in low light with the camera on tripod is not exactly easy as you have found out yourselves. If it was of any use, I would actually do just that.

timo

You know to be honest I haven't even tried 'Live view'.

I think it is a real pity that Nikon don't have the ultrasonic release that went with the D70 on the 'higher' models in the range. Using the Nikon remote cable thing is a pain in the .....

And they were pretty cheap (read disposable if your housing arrangements fail - I know you were using a ziplok bag there at one stage Timo ).

Think the switch and 'bulb' option is the way to go Timo.

Hey, I may try the gaffer tape idea on Alex's camera next week - it may hamper him somewhat initially but art is worth suffering for isn't it ?

Paul C

Hey, I may try the gaffer tape idea on Alex's camera next week - it may hamper him somewhat initially but art is worth suffering for isn't it ?

Don't get too cocky, Mr acryllic ports. Hope you have brought you polish (or save pole-camming until later in the trip)!

Alex

mmmmmm - yea, that is a bit scary....

Paul C

Pretty sure the exposure time is calculated just before the shot is taken - don't think that it is dynamic and calculated 'on the fly' during exposure.

I'd go with calculated before the shot for 2 reasons. 1) For my tests, shining the torch through the finder after hitting the shutter button didn't change the exposure time. 2) iTTL flash metering is all done prior to the actual shot being taken (albeit only a short time and after the shutter button is pressed). With film cameras, the flash exposure was metered off the film during the exposure but this is not done with digital so I doubt there is any metering off the sensor itself.